scotomas of age-related macular degeneration detected and

SCOTOMAS OF AGE-RELATED MACULARDEGENERATION DETECTED ANDCHARACTERIZED BY MEANS OF ANOVEL THREE-DIMENSIONALCOMPUTER-AUTOMATED VISUALFIELD TESTPAUL P. NAZEMI, MD,* WOLFGANG FINK, PHD,*† JENNIFER I. LIM, MD,*ALFREDO A. SADUN, MD, PHD*

Purpose: We used the recently devised three-dimensional computer-based thresholdAmsler grid test to acquire and identify typical patterns of visual field defects (scotomas)caused by age-related macular degeneration (AMD).

Methods: Patients with AMD traced on a computer touch screen the borders of thoseareas on an Amsler grid that were missing from their field of vision. Scotomas wererepeatedly outlined and recorded at different grid contrast levels. The resulting three-dimensional “hole” in the central 25° of the visual field was further characterized by itsslope, location, shape, and depth. The results were compared with fundus photographsand fluorescein angiograms.

Results: Twenty-five patients and 41 eyes were examined. The three-dimensionaldepictions consistently demonstrated central scotomas with “scallop”-shaped bordersand steplike patterns, with either steep slopes or a combination of steep and shallowslopes. The steep slopes corresponded to nonexudative AMD, while the shallow slopesindicated exudative AMD.

Conclusion: The three-dimensional computer-automated threshold Amsler grid testmay demonstrate characteristic scotoma patterns in patients with AMD that conform to therespective fluorescein angiograms. The test shows promise as an effective tool in accu-rately evaluating, characterizing, and monitoring scotomas in patients with AMD. It mayhave the potential as a screening tool for the early diagnosis of AMD.

RETINA 25:446–453, 2005

Age-related macular degeneration (AMD) is theleading cause of visual impairment in people

older than 65 years of age in the Western world.1Between 4 and 10 million people in the United States

are estimated to have AMD.2–7 The advanced formcan be choroidal neovascularization and geographicatrophy. Patients are defined as having the earliestform of nonexudative AMD if an eye has drusen of atleast 63 !m in size. In some of these patients andespecially in patients with large soft confluent drusenwith retinal pigment epithelium mottling, there is arisk for conversion to the exudative (neovascular)form. An estimated 80% of AMD patients have thenonneovascular form; however, the neovascular formmay account for almost 90% of the severe visual loss(20/200 or worse in both eyes) due to AMD.8

From *Doheny Eye Institute and Keck School of Medicine at theUniversity of Southern California, Los Angeles, California; and†California Institute of Technology, Pasadena, California.

The authors have financial/proprietary interest in materials usedin this study.

Reprint requests: Dr. Alfredo A. Sadun, Department of Ophthal-mology and Neurologic Surgery, Doheny Eye Institute and KeckSchool of Medicine at USC, 1450 San Pablo Street, DEI 5802, LosAngeles, CA 90033–4671; e-mail: [email protected]

446

Despite the uncertainty about the pathogenesis ofAMD, beneficial treatment exists, such as the use ofverteporfin in patients with predominantly classic subfo-veal choroidal neovascularization.9 With several treat-ment possibilities for neovascular AMD, early detectionand diagnosis may improve treatment outcome. To ad-dress this possibility of a screening method for the earlydetection of AMD, we first set out to characterize thevisual field defects caused by macular degeneration.

Visual field testing has advanced greatly since thedevelopment of the Amsler grid by Mark Amsler inthe 1920s. Testing for macular degeneration did notbecome routine until the dissemination of the Amslergrid in the middle of the 20th century. The Amslergrid was designed to specifically test visual field de-fects in the central 10°.10,11 An inexpensive and rapidmethod, the Amsler grid test utilizes a suprathresholdtarget to analyze the central 10°. It is good for detect-ing metamorphopsia but is not sensitive for the detec-tion of relative scotomas.11 The detection of relativescotomas was advanced by the development of thethreshold Amsler grid test by Wall and Sadun11 andWall and May.12 By using cross-polarizing filters tovary perceived luminance, threshold Amsler grid test-ing significantly increases the yield of scotoma detec-tion in the central 10°. Accordingly, it was found to bemuch more sensitive for shallow scotoma detection inpatients with maculopathies and optic neuropa-thies.11,12 Threshold Amsler grid testing of the central25° or more can now be easily and rapidly accom-plished with the recently introduced three-dimensionalcomputer-automated threshold Amsler grid test.13–16

This psychophysical test may be used to rapidly identifyand monitor patients with a variety of visual field defects.It has been shown to provide several advantages overconventional perimetry, primarily by adding a third (z)dimension in terms of contrast sensitivity to the x and ycoordinates of the visual fields and in characterizing anddistinguishing optic neuropathies.

The purpose of our study was thus to first retrievetypical signature patterns of AMD and then to estab-lish the size, shape, slope, and extent of the visual fielddefects using this three-dimensional computer-auto-mated visual field test. This study also aimed to fur-ther distinguish the visual field defects of AMD pa-tients with respect to exudative (i.e., neovascular)versus nonexudative (i.e., drusen or geographic atro-phy) forms of macular degeneration.

Patients and Methods

We recruited 25 patients from the Doheny EyeInstitute at the Keck School of Medicine of the Uni-versity of Southern California (Los Angeles). Patients

were recruited from the retina clinic directly or con-tacted and scheduled to participate in the study. Insti-tutional review board approval was obtained. Therewas no minimum requirement for visual acuity. Atotal of 41 eyes were tested using the three-dimen-sional computer-automated visual field test. The ex-amination employed the equipment introduced byFink et al13 and Fahimi et al14 and was performed in adesignated examination room. We utilized an IBMcompatible Pentium II PC running Windows 98 inconjunction with a 17-inch touch-sensitive computermonitor. The brightness level was well defined andgraded systematically throughout each study and be-tween the separate trials. The monitor was not recali-brated but was kept at the same brightness and con-trast settings both throughout each study and betweenseparate trials for consistency. Each patient was posi-tioned in front of the computer monitor. The angle ofthe visual field was determined by seating the patientat the fixed distance of 12 inches from the centralfixation marker on the computer screen (0° horizon-tally and 0° vertically from fixation). An eye coverwas used to occlude the eye that was not being exam-ined. Refractive correction was used with the patient’seyeglasses when necessary. The patients utilized theirspectacle (bifocal) correction during the test whenappropriate, and additional corrective lenses were notused.

An Amsler grid at a preselected gray scale (i.e.,contrast) level and preselected angular resolution wasdisplayed by the computerized test program. The pre-selected angular resolution for all patients was deter-mined by the distance between the monitor and thepatient and the Amsler grid spacing on the monitor. Itwas consistently set to 1° (standard Amsler grid) be-tween lines for each patient and study. This expandedto represent 60° ! 44° for the entire visual field test.The patient was first asked to focus on a changingfixation marker at the center of the grid. The fixationmarker did not change in brightness, independentfrom the gray scale level of the Amsler grid. Tosuppress the central Troxler effect and keep the pa-tient’s attention, the shape of the fixation marker wasregularly changed. Given the instability of fixation inpatients with central scotomas, the patient was in-structed to use the four corners of the rectangularmonitor as an additional reference frame for fixation.The patient was asked to mark the areas on the Amslergrid that were missing from his or her field of visionby tracing the border of this region with their finger onthe touch screen while maintaining fixation. The pa-tient’s response was recorded by the computerprogram.

447AMD CHARACTERIZED BY 3D AMSLER GRID TEST • NAZEMI ET AL

After completion of a particular gray scale level, thecontrast/gray scale level of the Amsler grid waschanged, and the patient had to perform the same taskas outlined above. This procedure was repeated for atotal of five different contrast levels, with one levelbeing the standard Amsler grid, in our case a whitegrid on a black background (i.e., 100% contrast dif-ference with respect to the background of the com-puter screen). The gray scale levels were presented inan ascending order. Areas of 0% contrast sensitivitywere then defined as the inability of a patient torecognize an Amsler grid displayed at 100% contrastdifference in those areas. Areas of 100% contrastsensitivity were defined as the ability of a patient torecognize an Amsler grid displayed at the lowestpreset contrast (i.e., the darkest grid). By changing thecontrast at which the Amsler grid was presented, hor-izontal “cuts” through the hill-of-vision at various“heights” (i.e., contrast sensitivity levels) were per-formed, which, in combination, resulted in an instan-taneous three-dimensional depiction of the hill-of-vi-sion (thus the name of the novel test).

Increasing degrees of contrast were simulated byrepeating this procedure at different preselected grayscale levels. The results were recorded and later dis-played as topographic contour rings by the computer-ized test program after each recording. A three-dimen-sional depiction of the central 25° of the visual fieldwas attained using these results, which could also bedisplayed as color shaded depictions in three dimen-sions that could be viewed or rotated to any perspec-tive. The results were then used to establish the loca-tion, extent, slope, depth, and shape of the scotomas inthe AMD patients. Each eye required a total of "5minutes to be tested.

Description of the Slope Measurement Method

The ratio between the loss of contrast sensitivityover degrees of visual field taken perpendicularlyto the closest contour rings expressed as a grade(%/degree) was initially computed to determine thesteepest slope of the scotoma. In the second method,the ratio between the loss of contrast sensitivity overdegrees of visual field taken perpendicularly to theleast compacted contour rings expressed as a grade(%/degree) was considered as a measure of the leaststeep (shallowest) side of the three-dimensional“hole.” In addition, the scotomas were characterizedusing the ratio of the area of the scotoma at the highestand lowest contrast levels. This was expressed as ascotoma area ratio (SAR), defined as the area of thescotoma at the highest contrast level divided by thearea of the scotoma at the lowest contrast level. Sco-

tomas with relatively steep slopes are characterized bySARs close to 1, whereas scotomas with predomi-nantly shallow slopes are characterized by lowerSARs.

Results

Forty-one eyes of 25 patients (17 women and 8men) with AMD underwent examination with thethree-dimensional computer-automated visual fieldtest. Visual acuities ranged from 20/60 to 20/200. Thenonexudative group included patients with focal hy-perpigmentation along with #5 drusen and confluenceof the soft large (#63 !m) drusen, and geographicatrophy. The three-dimensional depictions of visualfield loss associated with macular degeneration con-sistently demonstrated large central scotomas with“scallop”-shaped borders characterized by manychanges in the radius of curvature of the borders. Allof these patients demonstrated absolute scotomas onthe three-dimensional computer-automated visualfield test. The patterns revealed a characteristicsteplike pattern, with steep slopes or a combination ofsteep and shallow slopes (Figs. 1–3). The outlinedscotoma boundaries for each tested contrast level (to-tal of 5 gray scale levels) were concentrically ar-ranged, and the lines were overlapping and not cross-ing. The mean of the steepest slope (%/degrees ofvisual field) for the 41 eyes was 73.2%/degree (range,16%/degree to 80%/degree). The mean of the shal-lowest slope was 30.0%/degree (range, 5.6%/degreeto 80%/degree). All of the scotomas were absolute tothe highest gray scale Amsler grid (100% contrastdifference).

The steep slopes determined by the three-dimen-sional Amsler grid test corresponded to nonexudativeAMD, as determined by the fundus photographs andfluorescein angiograms. The shallow slopes were seenin patients with exudative AMD. Further, the area ofshallow slopes often corresponded to the area of neo-vascularization on the fundus color photographs andfluorescein angiograms. Figure 1 demonstrates an ab-solute central scotoma with a steplike pattern andcombination of steep and shallow slopes of the righteye. The area of laser photocoagulation in the subfo-veal region corresponds with the steepest slope on thethree-dimensional Amsler grid test (80%/degree).Subretinal fluid and adjacent subretinal blood are seennasal and superior to the laser scar, with a correspond-ing shallow slope (20%/degree). The marked scotomaarea measured 57 degree2 at the lowest tested grid to24 degrees2 at the highest tested grid. The SAR mea-sured 0.42. An active membrane in the right eye of apatient with classic AMD is seen in Figure 2. The shal-

448 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES ● 2005 ● VOLUME 25 ● NUMBER 4

lowest slope (8%/degree) seen temporally on thethree-dimensional Amsler grid corresponds to theslightly nasal location of the choroidal neovascularclassic component on the fundus photograph. The area ofthe scotoma measured 1,388 degree2 and 852 degree2 atthe area of the lowest and highest tested grids, respec-tively, and the SAR was 0.61. Early- and late-phaseangiograms in Figure 3 show leakage corresponding tochoroidal neovascularization in the subfoveal region ofthe left eye. A corresponding three-dimensional Amslergrid test reveals a shallow slope (8%/degree) and asteepest slope of 80%/degree. The area of the scotomawas 594 degree2 for the area at the lowest tested grid and274 degree2 for the area at the highest tested grid. TheSAR measured 0.46.

Discussion

AMD is the leading cause of visual impairment inthe elderly and the most common cause of permanentblindness in the developed world. As such, it repre-

sents a major public health problem. The usefulness ofautomated perimetry in the differential diagnosis of pa-tients with macular disease has been confirmed.17 How-ever, automated visual field testing often is not regularlyperformed on macular degeneration patients due to alack of resolution and fixation. There is a paucity ofliterature with detailed description of visual field defectscaused by macular degeneration. Unlike conventionalperimetry and campimetry that provide information per-taining to the borderline between seeing and nonseeingareas but do not provide the additional information in-herent in a region of visual depression, the computerizedthree-dimensional test displays the area of depression onthe z-axis in regards to changes in visual contrast sensi-tivity in addition to the area of depression in the x-axisand y-axis (relative to visual fixation). The three-dimensional information permits further distinctions be-tween certain visual disorders. Other important advan-tages of this technique over conventional perimetry havealready been described.13–16

Fig. 1. A, Fundus photographof the right eye shows an areaof laser photocoagulation in thesubfoveal region. Nasal and su-perior to this is subretinal fluidwith some adjacent subretinalblood. B, Laminar-phase fluo-rescein angiogram shows hy-perfluorescence correspondingto a laser scar with deeper cho-roidal vessels visible. Superiorto this are areas of speckledhyperfluorescence that showleakage on late frames of theangiogram (not shown). Na-sally, there is blocked fluores-cence due to hemorrhage. C,Gridlike depiction of three-di-mensional visual field in a pa-tient with age-related maculardegeneration that was re-corded by the three-dimen-sional computer-automatedthreshold Amsler grid test.The x-axis and y-axis denotethe horizontal and vertical co-ordinates of the visual field,respectively, in degrees with(0,0) being the center of fixa-tion. The z-axis denotes thecontrast sensitivity of the vi-sual field expressed in per-cent: 0% contrast sensitivitycorresponds to the standard(brightest) Amsler grid, and100% contrast sensitivity cor-responds to the Amsler griddisplayed at the lowest presetcontrast (i.e., the darkestgrid). T, temporal.


In our study, we set out to use three-dimensionalcomputer-based threshold Amsler grid test technologyto characterize the visual field defects caused by mac-ular degeneration. Most forms of automated perimetryhave proven most useful in the detection of glaucomaand other types of optic neuropathies. Improvementsin resolution and reliability have made computer-based testing increasingly more common as a usefuldiagnostic tool for detecting retinal disease. For ex-ample, scanning laser entoptic perimetry has recentlybeen shown to be an effective and inexpensive test forscreening retinal disease in visually asymptomatic pa-tients and glaucoma patients.18,19 Because the earlydiagnosis of exudative AMD may be more critical foreffective treatment, we tested the additional benefitsof greater sensitivity, higher resolution, better reliabil-ity, and additional information made possible by com-putation and consideration of the third dimension.

The present study was undertaken to first obtain andidentify distinctive signature patterns for defectscaused by AMD. Patients with previously known di-agnoses of AMD were evaluated. The results of thetest were as expected in terms of revealing a centrally

located, circular shaped visual field defect. Patientswith very variable fixation might outline differentareas at different contrast levels and may producevariable findings from session to session. The patientsin this study had difficulty identifying the centralfixation marker and were instructed to use the fourcorners of the monitor as a reference frame for fixa-tion. The success of this arrangement is demonstratedby the finding that the outlined scotoma boundaries foreach tested contrast level (total of 5 gray scale levels)were concentrically arranged and the lines were over-lapping and not crossing. This implies that the patientswere able to maintain adequate fixation. However,repeated testing within each of the five contrast levelswas not performed. This will be necessary to deter-mine the variability of the test within subjects. Thedepth, shape, and slope of the field defects were thencalculated. The patterns of visual field defects in pa-tients with AMD demonstrated a combination of steepslopes interrupted by stepwise shallow slopes. Thepatients in our study had later stages of AMD anddemonstrated absolute scotomas on the three-dimen-sional computer-automated visual field test. Patients

Fig. 2. A, Fundus photograph of the right eye shows subfoveal subretinal fluid, hard exudates, and drusen. Areas of subretinal hemorrhage arepresent. An area of choroidal neovascularization is visible as grayish green discoloration in the subfoveal region. The scotoma involves centralfixation. B, Early laminar–phase angiogram shows an area of subfoveal choroidal hyperfluorescence with some blocked fluorescence correspondingto blood. C, Late-phase angiogram of the right eye shows areas of hyperfluorescence as well as leakage in the subfoveal space. There are still someareas of blocked fluorescence corresponding to the blood in the right eye. D, Fully shaded depiction—equivalent to the gridlike display—of thethree-dimensional visual field in a patient with age-related macular degeneration that was recorded by the three-dimensional computer-automatedthreshold Amsler grid test (for explanation of axes, see Fig. 1C). The scotoma involves fixation. The temporal field to the center correlates with theslightly nasal location of the choroidal neovascular classic component. T, temporal.


with earlier stages would more likely have shown noscotoma at all or a relative scotoma. Such visual fielddeficits may escape detection altogether by the stan-dard Amsler grid. So far, it appears that the shape andslope patterns of the scotomas in AMD are unique forthis disease. However, in the absence of comparisonof these patterns with other macular diseases as wellas in the absence of testing across a larger sample,more studies may be called for to better characterizethe pattern of visual field loss in AMD patients. Thesteep slopes indicated nonexudative AMD, whereasthe shallow areas often pointed to areas of neovascu-larization as confirmed by color photographs and flu-orescein angiograms. Areas of steep slopes were alsoseen in exudative cases. In Figure 2, the slight nasallocation of the choroidal neovascularization on thefundus photograph demonstrates a correspondingshallow slope located (8%/degree) on the three-di-mensional Amsler grid test. However, the choroidalneovascularization also extends temporally as seen on

the fluorescein angiogram but has a steep border(80%/degree). A possible reason for this finding,which has not yet been clarified, is that the area ofgrayish green discoloration, seen nasally on the fun-dus photograph, may represent a more long-standinglesion, whereas the area extending temporally repre-sents a new lesion that has not caused pigmentarychanges.

The SARs were also calculated to better character-ize the scotomas. For example, both Figures 2 and 3demonstrate a shallow slope of 8%/degree and a steep-est slope of 80%/degree. However, there is a differ-ence in the relative frequency of occurrence of bothshallowest and steepest slopes in both figures, as dem-onstrated by their respective SARs. On the basis ofthese findings, it seems reasonable to suggest that thethree-dimensional computer-automated thresholdAmsler grid test may be used to frequently monitorpatients with nonexudative AMD for the developmentof choroidal neovascularization. The early diagnosis

Fig. 3. A, Early laminar–phase fluorescein angiogramshows a subfoveal area ofchoroidal neovascularizationwith surrounding blocked flu-orescence corresponding toblood and occult hyperfluo-rescence beyond the blood. B,Late-phase angiogram show-ing leakage corresponding tochoroidal neovascularizationin the subfoveal region of theleft eye. C, Gridlike depictionof the three-dimensional vi-sual field in a patient withage-related macular degener-ation that was recorded by thethree-dimensional computer-automated threshold Amslergrid test (for explanation ofaxes, see Fig. 1C). Horizontaland vertical lines extending tothe periphery represent arti-facts. T, temporal.


of exudative AMD and its prompt treatment may beuseful in mitigating further damage to the retina withresultant visual field loss. Clearly, this is a pilot study,and thorough longitudinal follow-up studies are re-quired to assess this hypothesis.

Our 21 patients were also tested with standardAmsler grid methodology, and the two tests werecompared for their ability to characterize absolutevisual field defects. Amsler grid testing did not accu-rately indicate the shape and extent of the scotoma. Apossible reason for this is that with standard Amslergrid testing, the patient may have difficulty maintain-ing fixation or charting the changes that may often beconfusing. In contrast, patient motivation and partici-pation in the three-dimensional computer-basedthreshold Amsler grid test is higher due to the periph-eral help with fixation and clear changes in boundariesthat occur at different contrast depths. This may resultin a more precise depiction of the visual field defect.Of course, the standard Amsler grid test serves only todetect the presence of an absolute scotoma, withoutany further information such as depth or slope. Inaddition, it is well known that Amsler grids mayunderestimate both the extent and the degree of sco-tomas as compared with SLO perimetry.20 However,this study did not compare the three-dimensional com-puter-based threshold Amsler grid test with conven-tional computerized visual field testing for severalreasons. First, we wished to determine whether thethree-dimensional Amsler grid test could help to refinethe scotoma determination as compared with the stan-dard Amsler grid. Second, we wished to determinefunctional correlates that may have been implicated byany features of the test that were specific to certainfundus findings. Hence, we sought to correlate thethree-dimensional Amsler grid test findings with thetopographic correlates on the fundus examination foreach patient. As we found, the steepness of the slopedoes seem to correlate with the fundus finding, unlikethe more nonspecific standard Amsler grid changes.Conventional computerized visual field testing is notvery reliable, such as for problems with fixation. Fu-ture studies could be performed comparing ourmethod with microperimetry. Furthermore, ourmethod is also relatively fast ($5 minutes/eye) andoffers additional information through three-dimen-sional depiction of scotomas and potential worldwideaccessibility over the Internet, as compared with mi-croperimetry mapping of the macular field andscotomas.

The visual field defects in patients with AMD usu-ally begin as relative scotomas and may progress tobecome absolute in the center over time. This is incontrast to the visual field defects in optic neuritis,

which often start out as areas of absolute scotoma thatover time often resolve to patches of relative scotoma.Therefore, AMD may be a more important diseasewith which to look for relative scotomas in the earlydetection of potentially treatable forms that may beimportant for the prevention of severe visual loss.

The three-dimensional computer-automated thresh-old Amsler grid test may also have the potential as ascreening tool for the early detection of AMD ingeneral. This was demonstrated as a relative scotomain a 43-year-old patient with a family history of AMDin both parents who presented with metamorphopsiaand no previous diagnosis of macular degeneration.As expected, given that the three-dimensional visualfield test was a more sensitive measure, a relativescotoma was detected that was not identified by thestandard Amsler grid. The scotoma was characteristicof our family of macular degeneration scotomas. Spe-cifically, the patient’s visual field examination dem-onstrated a centrally located relative scotoma with“scallop”-shaped borders, steplike pattern, and a steepslope (80%/degree). The three-dimensional depictionof the visual field loss in this patient may be sugges-tive of an early stage of nonexudative AMD. Hence,we believe that this computerized psychophysical testmay be an effective tool in the early diagnosis ofAMD. Its clinical utility will become even more ap-parent with the development of further treatments forAMD. The capability of the three-dimensional Amslergrid test to detect shallow or relative scotomas inaddition to absolute scotomas, combined with thecharacteristic pattern of visual field defects in AMDpatients, may enable the earlier detection of AMD.However, this study did not evaluate the use of the testas a screening tool, and further studies will be neededto address this possibility.

The three-dimensional computer-automated visualfield test shows promise as an effective tool in accu-rately evaluating, characterizing, and monitoring thevisual field defects in patients with AMD. Further, itmay have the potential as a screening tool for the earlydiagnosis of AMD.

Key words: three-dimensional computer-basedthreshold Amsler grid test, age-related macular degen-eration (AMD), Amsler grid, contrast sensitivity, ex-udative AMD, nonexudative AMD, perimetry, scoto-mas, visual fields, visual field testing.

References

1. Leibowitz HM, Krueger DE, Maunder LR, et al. The Fra-mingham Eye Study Monograph: an ophthalmological andepidemiological study of cataract, glaucoma, diabetic reti-nopathy, macular degeneration, and visual acuity in a general


population of 2631 adults, 1973–1975. Surv Ophthalmol1980;24:335–609.

2. Kahn HA, Moorhead HB. Statistics on Blindness in theModel Reporting Area 1969–1970. Washington, DC: USDepartment of Health, Education, and Welfare; 1973. DHEWpublication NIH73–427.

3. Klein R, Klein B, Linton K. Prevalence of age-related macu-lopathy: the Beaver Dam Eye Study. Ophthalmology 1992;99:933–943.

4. Kahn HA, Leibowitz HM, Ganley JP, et al. The FraminghamEye Study, I: outline and major prevalence findings. Am JEpidemiol 1977;106:17–32.

5. Sommer A, Tielsch JM, Katz J, et al. Racial differences in thecause-specific prevalence of blindness in East Baltimore.N Engl J Med 1991;325:1412–1417.

6. Klein B, Klein R. Cataracts and macular degeneration inolder Americans. Arch Ophthalmol 1982;100:571–573.

7. National Center for Health Statistics. Health, United States1989. Hyattsville, MD: Public Health Service; 1990.

8. Feris FL, Fine SL, Hyman L. Age-related macular degener-ation and blindness due to neovascular maculopathy. ArchOphthalmol 1984;102:1640–1642.

9. Henney JE. New therapy for macular degeneration. JAMA2000;283:2779.

10. Amsler M. Earliest symptoms of disease of the macula. Br JOphthalmol 1953;37:521–537.

11. Wall M, Sadun AA. Threshold Amsler grid testing: cross-polarizing lenses enhance yield. Arch Ophthalmol 1986;104:520–523.

12. Wall M, May DR. Threshold Amsler grid testing in macu-lopathies. Ophthalmology 1987;94:1126–1133.

13. Fink W, Hsieh AK, Sadun AA. Computer-automated 3-Dvisual field testing in distinguishing paracentral scotomas ofoptic neuritis vs. AION [ARVO abstract 1643]. Invest Oph-thalmol Vis Sci 2000;41:S311.

14. Fahimi A, Sadun AA, Fink W. Computer automated 3Dvisual field testing of scotomas in glaucoma [ARVO abstract796]. Invest Ophthalmol Vis Sci 2001;42:S149.

15. Fink W, Sadun AA. Novel 3D computerized thresholdAmsler grid test. Perimetry update 2002/2003. 15th Interna-tional Perimetric Society Meeting; June 2002; Stratford UponAvon, England.

16. Fink W, Sadun AA. 3D Computer-automated thresholdAmsler grid test. Journal of Biomedical Optics 2004;9:149–150.

17. Sobolewski P, Slomska J, Pienkowska-Machoy E, et al. Ap-plication of automated perimetry in diagnosis of maculardisease. Klin Oczna 1999;101:255–259.

18. Plummer DJ, Azen SP, Freeman WR. Scanning laser entopticperimetry for the screening of macular and peripheral retinaldisease. Arch Ophthalmol 2000;118:1205–1210.

19. Plummer DJ, Lopez A, Azen SP, et al. Correlation betweenstatic automated and scanning laser entopic perimetry innormal subjects and glaucoma patients. Ophthalmology2000;107:1693–1701.

20. Schuchard RA. Validity and interpretation of Amsler gridreports. Arch Ophthalmol 1993;111:776–780.


scotomas of age-related macular degeneration detected and

Documents